Artificial intelligence equals radiologist performance in prostate MRI cancer detection across diverse datasets

An AI system achieved AUCs of 0.86–0.89 for detecting clinically significant prostate cancer on MRI across three international datasets, meeting criteria for diagnostic interchangeability with radiologists.

• The retrospective study evaluated 22,445 prostate MRI examinations from 21,252 patients across 14 countries, with external validation in 12 countries.
• Performance was consistent across patient subgroups by age, cancer risk, and ethnicity, but varied by clinical site and image quality.

Automated summary

RadPigeon summaries are original and for information only. They are not clinical advice.